Providers of television services are interested in viewership information about what programs and channels are being watched by television viewers. The providers of television services may want to monitor viewing habits to gather information for use in making decisions about programming. Additionally, viewers may want to know what is popular with other viewers. Such services require that the viewership information is updated quickly. However, conventional data collection systems for determining viewership can be slow to update, processor intensive, bandwidth intensive (adversely impacting performance), or any combination thereof. Additionally, such data collections systems may not be scalable as the number of viewers increases. Hence, there is a need for an improved system and method of determining viewership information.